• Sediment movement through riparian vegetation under simulated rainfall and overland flow

      Pearce, R. A.; Trlica, M. J.; Leininger, W. C.; Mergen, D. E.; Frasier, G. (Society for Range Management, 1998-05-01)
      A 2-year rainfall simulation study was conducted to evaluate the effectiveness of riparian vegetation to reduce sediment movement. Three vegetation height treatments [clipped to the soil surface, 10 cm height, and undisturbed (unclipped)] were evaluated in 2 montane riparian vegetation communities in northern Colorado. One community was a tufted hairgrass (Deschampsia caespitosa (L.) Beauv.), cinquefoil (Potentilla gracilis Dougl. ex Hook), Kentucky bluegrass (Poa pratensis L.), and sedge (Carex spp.) association. The other community was dominated by beaked sedge (Carex rostrata Stokes) and water sedge (Carex aquatilis Wahl.). Water was sprayed on plots (3 m X 10 cm) at a rate of 60 mm hour-1 with a rainfall simulator, while overland flow containing sediment was introduced at the upper end of the plots at a rate of 25 mm hour-1. Two sediment sources were evaluated. In the first year's studies the sediment was derived from an upland soil. The second year a fine silica sediment was used. Thirty kg of sediment was added to each plot. The first of 2 experiments was concerned with movement of sand particles greater than 200 micrometers. The second experiment was designed to evaluate the quantity of 5 particle size classes (2-10, 10-30, 30-50, 50-100, and 100-200 micrometers) contained in sediment traps at 60, 120, and 180 cm downslope from the upslope border of the simulator plots. Results of the first experiment showed a significant increase in sand movement downslope when vegetation was clipped to the soil surface compared with undisturbed vegetation. In the second experiment, most significant differences in movement for finer particles occurred in the 2-10 micrometers and 10-30 micrometers particle ranges. A smaller percentage of particles in the 2-10 micrometers range was present in sediment traps at all 3 distances downslope when vegetation was clipped to the soil surface, as these smaller particles stayed in suspension. Increased vegetation height resulted in a significantly smaller percentage of the 10-30 micrometers particle size range present at 120 cm distance. This study showed that additional variables (% surface vegetation cover, aboveground biomass, % shrubs, surface roughness coefficient, soil texture of introduced sediment, % bare ground, distance downslope, vegetation density, grass spp., and sedge spp.), besides vegetation height, influenced sediment movement. Land managers should understand that when they manage ecosystems for a single factor, such as vegetation height, they cannot address complex issues such as sediment particle detachment, movement and filtration.
    • Technical note: Predicting the components of aerial biomass of fourwing saltbush from shrub height and volume

      Thomson, E. F.; Mirza, S. N.; Afzal, J. (Society for Range Management, 1998-05-01)
      Shrub height and crown diameter are useful non-destructive measures of shrub growth, but precise yields of aerial biomass require destructive methods which are unsatisfactory in studies on perennial shrubs. We developed simple regression models to predict components of aerial biomass from the height, crown diameter and volume of 27 unbrowsed shrubs of fourwing saltbush (Atriplex canescens). The shrubs, ranging in height from 15 to 110 cm, were cut at ground level and manually separated into forage (leaves) and woody material. Samples were oven-dried. Shrub height and volume were sufficiently precise for predicting components of aerial biomass using exponential and linear regression models, respectively. The precision of these non-destructive measures applied under field conditions to unbrowsed shrubs should be confirmed on browsed shrubs.
    • Thurber needlegrass: Seasonal defoliation effects on forage quantity and quality

      Ganskopp, D. (Society for Range Management, 1998-05-01)
      Although Thurber needlegrass (Stipa thurberiana Piper) is an important component of Palouse, sagebrush:steppe, and pine:forest rangelands, little is known of its qualitative and quantitative responses to defoliation. At 14-day intervals one of 7 cohorts of Thurber needlegrass plants was defoliated to a 2.5-cm stubble to describe initial growth rates, determine defoliation effects on subsequent regrowth accumulations, relate regrowth potential to available soil moisture, and determine the nutritional value of initial growth and regrowth for livestock. The study was conducted in 1985 and 1986 with a different group of plants used each year. Although crop-year precipitation for the 1985-86 treatment years was 77 and 111%, respectively, of the long term mean (25.2 cm), growth rates of tussocks were similar between years (P > 0.05). Seasonal yield of regrowth varied between years, however, and was well correlated (r2 = 0.76 to 0.80 P < 0.05) with soil moisture content when treatments were applied. Among 7 defoliation dates (24 April-17 July) only the first 5 yielded regrowth in 1985, and all produced regrowth in 1986. Among treatments regrowth averaged 22% of total herbage yield in 1985 and 50% of total yield in 1986. In both years total herbage accumulations were most suppressed (47-63% reduction) by defoliation during the early-boot stage of phenology. In 1985 when conditions were drier, any defoliation before mid-June depressed (P < 0.05) total herbage yield. Crude protein (CP) of needlegrass herbage was high (19-22%) when growth began in April but declined (P < 0.05) to marginal levels for cattle (6.7-7.7%) by mid-July. Regrowth harvested on 31 July ranged from 7 to 9% CP for the earliest (24 April) treatments and as high as 17% for the latest (17 July). Although Thurber needlegrass can produce highly nutritious regrowth for late-season use, managers face diminishing levels of regrowth as the initial cropping date is delayed later into the growing season. Managers contemplating 2-crop grazing regimes for Thurber needlegrass should base scheduling on plant phenology, soil moisture considerations and historic use rather than specific calendar dates. Further work is needed, however, to definitively determine Thurber needlegrass responses to long-term manipulative grazing regimes.
    • Viewpoint: The present status and future prospects of squirreltail research

      Jones, T. A. (Society for Range Management, 1998-05-01)
      Squirreltail's [Elymus elymoides (Raf.) Swezey + Sitanion hystrix (Nutt.) J.G. Smith] ready germination, rapid reproductive maturity, capacity for cool-temperature growth, self-pollinated mating system, excellent seed dispersal mechanisms, fire tolerance, and genetic diversity make it a promising candidate for assisting ecological restoration of rangelands dominated by exotic weedy annual grasses such as medusahead wildrye [Taeniatherum caput-medusae (L.) Nevski] and cheatgrass (Bromus tectorum L.). Squirreltail is a short-lived perennial and generally early seral in successional status. It comprises a complex of several subspecies whose ecological amplitudes are poorly understood. Wildfire or prescribed burning may provide opportunities for seeding squirreltail or augmenting existing populations. Grazing deferment is important for a successful transition from an annual to a perennial-dominated grassland. Reduction in frequency of annuals may facilitate natural or artificial establishment of desirable mid- or late-seral grasses, forbs, and shrubs. Currently, squirreltail seed supplies originate from wildland harvests. Reduced cost, dependable supply, and improved quality of seed will require development of efficient commercial seed production practices. Experience in restoration may reveal the suitability of squirreltail plant material for assisted succession as well as expose its weaknesses. Such information will allow researchers to improve plant materials and methods for increased future success.
    • Sheep production on medic and weedy pasture in semi-arid Morocco

      Tiedeman, J.; Boulanouar, B.; Christiansen, S.; Derkaoui, M. (Society for Range Management, 1998-05-01)
      Most farms in the semi-arid region of Morocco are mixed cereal-livestock producers. Livestock, dominated by sheep, account for more than one third of farm income but production is far below potential. Low quality cereal straw and weeds from fallow land are the main sources of forage, but provide inadequate nutrition. Fall planted wheat is usually grown in rotation following a year of weedy fallow. The replacement of the weedy fallow by sowing a self regenerating annual Medicago spp., a system known as ley farming in Australia, was tested in Morocco to determine if it could increase livestock production. Ewe and lamb liveweight gain, wool yield, and forage standing crop during grazing were compared between medic and weedy fallow pastures. Large significant differences (P < 0.05) in forage standing crop and both lamb and ewe liveweight gain occurred in 1990/91 but no differences occurred the previous 2 years when medic was initially sown nor the following 2 years during severe drought. In 1991 lambs gained 67% and ewes gained 60% more liveweight, plus wool yield was 23% higher on medic as compared to weedy fallow. In addition, 3,000 kg ha-4 of forage remained after grazing medic as compared to 568 kg ha-1 after weedy fallow.
    • Cover components on long-term seasonal sheep grazing treatments in three-tip sagebrush steppe

      Bork, E. W.; West, N. E.; Walker, J. W. (Society for Range Management, 1998-05-01)
      The effects of fall and spring sheep use on cover components and recovery following a change in seasonality of grazing practices, were studied within long-term grazing treatments of three-tip sagebrush (Artemisia tripartita Rydb.) steppe on the U.S. Sheep Experiment Station near Dubois, Ida. Few significant differences existed among treatments within the litter, moss, lichen, and soil components, but several differences in vegetational cover categories occurred. More live shrub and annual grass cover were observed in the long-term (since 1924) and new spring (since 1950) treatments than in the long-term fall (since 1924), new fall (since 1950), old exclosure (since 1940), and new exclosure (since 1950) (P < 0.01). More perennial grass and forb cover, and less dead shrub cover existed in fall-grazed treatments (P < 0.01). The new fall- grazed treatment previously grazed in the spring failed to reach a more uniform mixture of perennial growth forms after 46 years such as was evident in the long-term fall, which suggests low resilience following spring grazing. The exclosure which was heavily spring and fall-grazed prior to 1950 had even less perennial forb cover than the new fall treatment, indicating that the cessation of sheep grazing did not promote herb recovery any better than continued fall use. The direct impact of sheep herbivory and its indirect effects on the competitive relationships among major plants appear to have affected the cover of sagebrush steppe components at this study site.
    • Intraspecific competition in honey mesquite: Leaf and whole plant responses

      Ansley, R. J.; Trevino, B. A.; Jacoby, P. W. (Society for Range Management, 1998-05-01)
      Leaf and whole plant responses of honey mesquite (Prosopis glandulosa Torr.) to intraspecific competition were compared under low (LD) or high (HD) stand density in a semi-arid region of north Texas. The HD trees occurred within a stand of 300 trees ha-1. The LD trees occurred in areas of the dense stand that were thinned to 80 trees ha-1 with no neighbors within 10 m of study trees. Tree size was similar in each treatment at study initiation. Five years after thinning, tree height, canopy volume, basal stem diameter, leaf area, and leaf area index were significantly greater in LD than HD trees. No differences in leaf predawn water potential, stomatal conductance, and photosynthesis were found between LD and HD trees during growing seasons 4 or 6 years after study initiation. Results indicate resources necessary for growth of individual mesquite plants were limiting under increased stand density and suggest the occurrence of intraspecific competition. Limitations were manifest at the whole plant level via modification of tree size and leaf area per tree, and not through adjustment of leaf physiological processes. The limiting factor appeared to be soil water. Daily water loss tree-1 was 2.5 to 4 times greater in LD than HD trees, and ranged from 119 to 205 kg and 46 to 59 kg in LD and HD trees, respectively. Projected daily water loss by mesquite at the stand level was similar between treatments, however, and ranged from 9,500 to 17,700 kg ha-1.
    • Limpograss and hymenachne grown on flatwoods range pond margins

      Kalmbacher, R.; Mullahey, J.; Hill, K. (Society for Range Management, 1998-05-01)
      Limpograss (Hemarthria altissima [Poir] Stapf and C.E. Hubb) and hymenachne (Hymenachne amplexicaulis [Rudge] Nees) may reduce weight loss of cows grazing Florida range from September to March. These grasses were grown on maidencane (Panicum hemitomon Schult) pond margins and were evaluated as stockpiled forage (ungrazed 6-10 months) at 2 locations over 4 years. Floralta limpograss received 0 or 3,000 kg dolomite ha (2 whole plots) and N-P-K fertilizer (5 subplots): 50-25-50, 50-25-0, 50-0-50, 50-0-0, 0-0-0 kg/ha. Hymenachne was grown without dolomite, N, P, or K. Hymenachne failed to establish at Ona in central Florida, but persisted for 1 year at Immokalee near the Everglades where dry matter production in October to January was 1,540, 2,160, and 2,910 kg/ha at 35, 70, and 105 days after N fertilization, respectively. Crude protein (56 g/kg) was highest at 70 days and IVOMD (47.4%) was highest at 105 days. Limpograss established without dolomite, N, P, or K fertilization, and forage available for winter grazing often exceeded 7,000 kg/ha. Application of 50 kg N/ha to stockpiled limpograss increased yield (compared to no N) in 1 of 4 years at Ona and in both years at Immokalee. Applying N to stockpiled limpograss always increased crude protein and IVOMD above that of grass receiving no N, but increases were slight (10 g crude protein/kg). Crude protein seldom exceeded 50 g/kg with 50 kg N/ha applied in late August at Ona or in October at Immokalee. In vitro organic matter digestion often exceeded 45%, which could help limit weight loss of cows grazing range in winter. Neither grass was observed to be invasive, as growth was confined to plots after 5 and 8 years at Immokalee and Ona, respectively.
    • Bison grazing patterns on seasonally burned tallgrass prairie

      Coppedge, B. R.; Shaw, J. H. (Society for Range Management, 1998-05-01)
      Patterns of bison (Bison bison L.) grazing were examined in a 2-year study on a tallgrass prairie site in Oklahoma subjected to a seasonally and spatially variable burning regime. Mixed groups of bison, composed of cows, yearlings, calves, and young (< 5 years of age) bulls, comprised 90% of the study population and showed selectivity by using burned areas significantly more than expected 23% of the time. Mixed groups avoided unburned areas 63% of the time. In contrast, bull groups of mature bulls > 5 years of age selected unburned areas for grazing 29% of the time and burned areas only 4% of the time. Temporal patterns in bison grazing were evident; selective use of burns persisted for only a short period during the first post-fire growing season, after which burns were grazed in proportion to availability and then selectively avoided as bison shifted grazing efforts to newer burns. Regression analysis verified that bison grazing was negatively related to burn age. Regression also showed that grazing patterns were positively related to burn patch size. Although burn types varied significantly in biomass and overall vegetative composition, bison exhibited only limited preference for any burn type, choosing those with higher relative cover of annual Bromus spp. and sedges. It appears that bison select recently burned areas with relatively low graminoid biomass for grazing, presumably choosing these areas based on forage quality rather than quantity.
    • Environment and seedling age influence mesquite response to epicotyl removal

      Tischler, C. R.; Polley, H. W.; Johnson, H. B.; Mayeux, H. S. (Society for Range Management, 1998-05-01)
      Herbivory by small mammals is a major factor controlling survival of honey mesquite (Prosopis glandulosa Torr. var. glandulosa) seedlings. Clipping below the cotyledons is lethal; removal of the epicotyl may not be lethal but can severely limit seedling growth. Seedlings of other woody species sometimes compensate for epicotyl removal by prolonging the life of cotyledons. Also, projected future increases in atmospheric CO2 concentration could influence survival and growth after epicotyl removal. Objectives of this study were to determine effects of epicotyl removal at various seedling ages, atmospheric CO2 concentrations, and soil fertility, on (1) seedling survival, (2) cotyledonary leaf longevity, and (3) shoot and root growth of young seedlings. Mesquite seedlings were grown at 350, 700, and 1,000 microliters liter-1 atmospheric CO2 concentration in nutrient poor and nutrient rich soils. All ages of seedlings survived epicotyl removal. Cotyledonary leaf fresh mass and chlorophyll content were higher in plants where epicotyls were clipped. Root and shoot mass of both clipped and unclipped plants generally increased at higher CO2 concentrations when mineral nutrition was adequate, but responded less to CO2 when soil fertility was low. Responses to epicotyl clipping in mesquite seedlings are complex, being strongly influenced by soil fertility, atmospheric C)2 concentration, seedling age at clipping, and interactions between these factors.
    • Evaluating expert knowledge: plant species responses to cattle grazing and fire

      Rodríguez Iglesias, R. M.; Kothmann, M. M. (Society for Range Management, 1998-05-01)
      Expert judgment, standardized in a meaningful format, can be used to identify research/survey needs and to characterize areas of (dis)agreement in species responses, associated traits, and factors affecting responses. Feasible methods are needed to facilitate the evaluation of expertise in a complex domain characterized by moderate to low learnability. Specific objectives for this study were 1) to evaluate agreement among experts on range plant species behavior and 2) to develop an agreement-based classification method for plant species responses. Declarative information at landscape scale was elicited from 7 role-suggested experts on expected responses to cattle grazing (none, moderate, or heavy) and fire (absent, applied in late summer or fall, or applied in late winter or spring) of 198 plant species from the Edwards Plateau (Texas). Trends were requested to be assessed in a 3-level ordinal scale (decrease, unaffected, increase). Kappa statistics (pair-wise and multi-rater versions) and log-linear models were used to evaluate agreement. A procedure based upon cumulative probability distributions of possible rating combinations was developed to classify plant species while accounting for agreement. A total of 4,584 opinions (cattle grazing: 2,959; fire: 1,625) was elicited and analyzed. Low to moderate agreement was observed. Average pair-wise kappa statistics ranged from 0.07 to 0.39; multiple-rater kappa coefficients ranged from -0.17 to 0.53. Log-linear analyses were consistent with those estimations: agreement beyond chance or baseline association between ratings (P < 0.05) was observed in 62 out of 114 possible pair-wise cases. Non-homogeneous marginal distributions of opinion were an important source of disagreement. Experts performed beyond chance expectations in all scenarios but agreement was better (and pattern of agreement more consistent) when scenarios were most familiar to the experts (e.g., heavy grazing and winter/spring burning). Almost 80% of species was classified beyond chance (P < 0.15) in grazing scenarios in contrast to only 40 to 60% in fire scenarios. This resulted from less agreement among experts but also from apparent lack of knowledge. The procedure developed to classify plant species provides an objective criterion for evaluating agreement in an ordinal scale. Graphical representations facilitate understanding of relationships between the number of expert sources and their ability to distinguish among classes for a pre-defined confidence level.
    • Diets of 3 cattle breeds on Chihuahuan Desert rangeland

      de Alba Becerra, R.; Winder, J.; Holechek, J. L.; Cardenas, M. (Society for Range Management, 1998-05-01)
      Diet botanical composition, fecal nitrogen percent, and fecal phosphorus percent were determined seasonally during 1991 and 1992 for 3 cattle breeds (Barzona, Brangus, Beefmaster) grazing late-seral Chihuahuan Desert rangeland in southcentral New Mexico. These 3 cattle breeds are considered to be well adapted to harsh environments. Cattle breed main effect was non-significant (P > 0.05) for diet botanical composition. However, season main effects (P < 0.05) did occur for some diet botanical composition components. Total grasses in cattle diets were highest in January and lowest in June. Dropseeds (Sporobolus sp.), black grama (Bouteloua eriopoda Torr.), and threeawns (Aristida sp.) were the primary grasses consumed by cattle. Forb consumption was highest in June lowest in January. Honey mesquite (Prosopis glandulosa Torr.) consumption by cattle was highest in August and lowest in January. It was the primary shrub in cattle diets. Breed X season interactions (P < 0.05) occurred for a few diet botanical composition components, but the small magnitude of the values and lack of consistency prevented drawing definite management implications. Fecal nitrogen values showed differences (P < 0.05) among breeds in some seasons. However due to lack of consistency no definite conclusions could be drawn regarding superiority of 1 breed compared to another in diet nutritional quality. Both fecal nitrogen and phosphorus concentrations showed cattle diets to be lowest in quality in winter and highest in summer. From a practical standpoint, this study showed no definite advantage of any breed studied in diet botanical composition or in diet quality.
    • Nesting habitat selection by sage grouse in south-central Washington

      Sveum, C. M.; Edge, W. D.; Crawford, J. A. (Society for Range Management, 1998-05-01)
      To characterize western sage grouse (Centrocercus urophasianus phaios Bonaparte) nesting habitat in sagebrush-steppe habitat in Washington, we initiated a study on the Yakima Training Center to determine nesting habitat characteristics and whether these characteristics differed between successful and depredated nests. Most nests (71%) were in big sagebrush (Artemesia tridentata Nutt.)/bunchgrass communities. Nest habitat was characterized by greater shrub cover, shrub height, vertical cover height, residual cover, and litter than at random locations. Successful 1-m2 nest sites within big sagebrush/bunchgrass in 1992 had less shrub cover (51%) and shrub height (64 cm) than depredated nest sites (70% and 90 cm, respectively). Successful 77-m2 nest areas in big sage-brush/bunchgrass in 1993 had more tall grass (greater than or equal to 18 cm) than depredated nest areas. Management that protects the big sage-brush/bunchgrass community is essential for maintaining nesting habitat for sage grouse.
    • Relationships between physical and chemical characteristics of 3 Sandhills grasses

      Northup, B. K.; Nichols, J. T. (Society for Range Management, 1998-05-01)
      Physical and chemical traits of grass tillers can be strongly correlated. Understanding such patterns would help define physiological development of tillers and changes in quality of forage in Sandhills grasses. Physical and chemical traits were quantified for sand bluestem (Andropogon hallii Hack.), prairie sandreed [Calamovilfa longifolia (Hook.) Scribn.], and little bluestem [Andropogon scoparius (Michx.)] on 3 sites at 4 times (mid-June, July, August, and October) during the 1990 and 1991 growing seasons. Thirty tillers were identified along two, 50-m transects (30 tillers/species/transect) within each site and tiller growth stage, length, and erectness determined. Tiller weight was defined from plants collected within 20 quadrats/site. Protein content, in vitro dry-matter digestibility (IVDMD), hemicellulose, total cell wall, acid detergent fiber (ADF), lignin, ash, total chlorophyll, and nonstructural carbohydrates (TNC) were determined on plant materials representing the dominant growth stages. Relationships among traits of the 3 species were determined by Spearman's rank correlation, and among linear combinations of sets of chemical and physical traits by canonical correlation analysis. Tiller length, weight, and growth stage were positively correlated (P < 0.05) and increased with length of growing season. Crude protein, digestibility, hemicellulose and chlorophyll were positively correlated and declined, but negatively correlated with lignin and ash. Significant (P < 0.05) correlations between the first canonical variates indicated a strong relationship between tiller maturity/architectural development (physical canonical variate) and forage quality (chemical canonical variate) was present, and large portions of variance in the original variables was defined. Results of this study defined large-scale multi-dimensional relationships between declining forage quality and increasing tiller maturity/architectural development, previously noted in many univariate analyses of limited sets of characteristics.
    • Runoff from simulated rainfall in 2 montane riparian communities

      Frasier, G. W.; Trlica, M. J.; Leininger, W. C.; Pearce, R. A.; Fernald, A. (Society for Range Management, 1998-05-01)
      Riparian ecosystems are the final terrestrial zone before runoff water enters a stream. They provide the last opportunity to decrease non-point source pollution delivery to streams by removing sediments from overland water flow from uplands and roads. To quantify processes of sediment transport, filtration and deposition, it is necessary to determine runoff characteristics for the area. A rotating boom rainfall simulator was used to evaluate the effects of 3 vegetation height treatments (control, 10-cm stubble height, and clipped to the soil surface) in 2 montane riparian plant communities (grass and sedge) on runoff characteristics. Each rainfall simulation event consisted of 2 phases, a dry run of about 60 min followed by a wet run approximately 30 min later. There were no differences in time to runoff initiation for either dry or wet runs that could be attributed to vegetation height treatments for either plant community. It usually required more time for runoff to be initiated in the sedge community compared to the grass community. Generally, there were lower equilibrium runoff percentages from dry runs in the sedge community compared with the grass community. These differences were less during wet runs. Several runoff parameters had characteristics of runoff from water repellent soils. The organic layer on the soil surface exhibited signs of water repellency that reduced the water infiltration rate during the initial stages of a rainfall simulation. These results indicate that runoff and infiltration processes in the surface organic horizon of riparian zones may not respond in the classical manner. This characteristic has important implications if criteria developed in areas with less organic matter on the soil surface are used to manage overland flow in the zone. Additional studies are needed to fully describe infiltration and runoff processes in riparian plant communities.
    • Sediment filtration in a montane riparian zone under simulated rainfall

      Pearce, R. A.; Frasier, G. W.; Trlica, M. J.; Leininger, W. C.; Stednick, J. D.; Smith, J. L. (Society for Range Management, 1998-05-01)
      A 2 year study was conducted to evaluate the effectiveness of riparian vegetation to filter sediment from overland water flow. Three vegetation height treatments: clipped to the soil surface, clipped to a 10 cm height, and undisturbed were evaluated in 2 montane riparian vegetation communities (grass and sedge) in northern Colorado. Water was sprayed on 2 macro-plots (3 m X 10 m) and 2 micro-plots (0.6 m X 2 m) simultaneously at a rate of 60 mm hr-1 with a rotating boom rainfall simulator. Overland flow containing sediment was introduced at the upper end of the plots at a rate of 25 mm hr-1 to simulate runoff and sediment transport from an upland area. Two sediment sources were used, a sandy loam soil and a ground silica sediment (loam). Thirty kg of sediment were added to each macro-plot and 1.2 kg of sediment were introduced to each micro-plot (10 Mg ha-1). Sediment yields, at the downslope end of the plot, were greater when the finer silica sediment was introduced into overland flow as compared with sediment derived from the sandy loam soil. As expected the small micro-plots yielded more sediment and were often more sensitive to community and treatment differences than larger plots. We believe this resulted from the shorter travel distance. However, sediment filtration treatment effects were usually similar for both plot sizes. Sediment yields, measured at the outlet of the plots, did not decrease, or increase, as vegetation heights increased. Accurate prediction of sediment filtration from shallow flow in riparian zones required consideration of a combination of vegetation and soil surface characteristics.